Enhancing Disaster Awareness using BERT-Based Analysis of Twitter Chat

Disaster management and response are critical for minimizing the impact of natural and man-made disasters. Social media platforms, particularly Twitter, have emerged as valuable sources of real-time information during disasters. This paper presents a novel approach to enhance disaster awareness through the analysis of Twitter chat using Bidirectional Encoder Representations from Transformers (BERT). BERT is a powerful natural language processing model that has shown remarkable performance in various text analysis tasks. This work illustrates BERT for analyzing tweets related to disasters to improve situational awareness. This work uses disaster-related tweets dataset and preprocess them for analysis and and the data is fine-tuned using pre-trained BERT model. This research demonstrate that the BERT-based approach outperforms traditional methods like BILSTM-Glove with accuracy of 84.25% and BERT-BILSTM-CNN with accuracy of 95% in classifying disaster-related tweets.